Rate constants for the reactions of AsIII with FAC, NH2Cl, NHCl2, and O3, were measured by various techniques selected according to reactant characteristics and reaction rates.. The tren
Trang 1Kinetics and Mechanistic Aspects
of As(III) Oxidation by Aqueous
Chlorine, Chloramines, and Ozone:
Relevance to Drinking Water
Treatment
M I C H A E L C D O D D ,† N G O C D U Y V U ,‡
A D R I A N A M M A N N ,† V A N C H I E U L E ,‡
R E I N H A R D K I S S N E R ,§
H U N G V I E T P H A M ,‡ T H E H A C A O ,‡
M I C H A E L B E R G ,† A N D
U R S V O N G U N T E N *, †
Swiss Federal Institute of Aquatic Science and Technology
(EAWAG), 8600 Duebendorf, Switzerland, Center for
Environmental Technology and Sustainable Development
(CETASD), Hanoi University of Science, Nguyen Trai Street
334, Hanoi, Vietnam, and Laboratory of Inorganic Chemistry,
ETH Zurich, 8093 Zurich, Switzerland
Kinetics and mechanisms of As(III) oxidation by free
available chlorine (FACsthe sum of HOCl and OCl-),
ozone (O3), and monochloramine (NH2Cl) were investigated
in buffered reagent solutions Each reaction was found
to be first order in oxidant and in As(III), with 1:1 stoichiometry.
FAC-As(III) and O3-As(III) reactions were extremely
fast, with pH-dependent, apparent second-order rate
constants, k ′′app, of 2.6 ((0.1) × 105M-1s-1and 1.5 ((0.1)
× 106M-1s-1at pH 7, whereas the NH2Cl-As(III)
reaction was relatively slow (k ′′app) 4.3 ((1.7) × 10-1
M-1s-1at pH 7) Experiments conducted in real water
samples spiked with 50 µg/L As(III) (6.7 × 10-7M) showed
that a 0.1 mg/L Cl2(1.4 × 10-6M) dose as FAC was
sufficient to achieve depletion of As(III) to <1 µg/L As(III)
within 10 s of oxidant addition to waters containing
negligible NH3 concentrations and DOC concentrations
<2 mg-C/L Even in a water containing 1 mg-N/L (7.1 × 10-5
M) as NH3, >75% As(III) oxidation could be achieved
within 10 s of dosing 1-2 mg/L Cl2(1.4-2.8 × 10-5M) as
FAC As(III) residuals remaining in NH3-containing waters
10 s after dosing FAC were slowly oxidized (t1/2g 4 h) in the
presence of NH2Cl formed by the FAC-NH3reaction.
Ozonation was sufficient to yield >99% depletion of 50
µg/L As(III) within 10 s of dosing 0.25 mg/L O3(5.2 × 10-6
M) to real waters containing <2 mg-C/L of DOC, while
0.8 mg/L O3(1.7 × 10-5M) was sufficient for a water containing
5.4 mg-C/L of DOC NH3had negligible effect on the
efficiency of As(III) oxidation by O3, due to the slow kinetics
of the O3-NH3reaction at circumneutral pH
Time-resolved measurements of As(III) loss during chlorination
and ozonation of real waters were accurately modeled using the rate constants determined in this investigation.
Introduction
Arsenic is a common contaminant of groundwater resources
around the world (1-4) Soluble inorganic arsenic occurs in
surface waters and groundwaters primarily as a combination
of arsenous acid (As(III)) and arsenic acid (As(V)) (1) The
former state predominates under anoxic conditions (e.g., in oxygen-limited groundwaters), and the latter under oxic
conditions (1), although As(III) can exist as a meta-stable
species even in oxygen-rich environments, due to the slow
kinetics of its oxidation by oxygen (1) Generally, dissolved arsenic occurs in groundwaters at concentrations <5 µg/L (1, 2, 5) However, in certain regions, including the western United States (2, 6) and southern Asia (3, 4), groundwaters
utilized for drinking water often contain arsenic
concentra-tions in substantial excess of the 10 µg/L guideline value recommended by the WHO (5) and adopted as a regulatory limit by the EU (7) and USEPA (8).
When sufficient infrastructure is available, aqueous
arsenic concentrations can be lowered to e 10 µg/L by a
variety of conventional drinking water treatment methods, though many of these methods remove As(III) substantially
less efficiently than As(V) (9) In cases for which As(total) is
constituted in large part by As(III), arsenic removal by such methods can be improved by preoxidizing As(III) to As(V)
(9-11) However, oxidant-scavenging matrix constituents,
such as dissolved organic matter (DOM) and NH3- which can be present at high concentrations within reduced,
As-(III)-laden groundwaters (3, 12, 13), may impair As(III)
oxi-dation efficiency by competing with As(III) for available
oxidant (10) Quantitative knowledge of the rate constants
and mechanisms governing oxidation of As(III) by common drinking water oxidants would greatly facilitate modeling and optimization of As(III) oxidation processes for treatment
of such waters
Apparent second-order rate constants, k′′app, were mea-sured for oxidation of As(III) by FAC, NH2Cl, and O3within the pH range 2 to 11, to permit evaluation of pH-dependencies for each reaction Stoichiometries and reaction orders were also measured, to facilitate identification of probable oxida-tion mechanisms Addioxida-tional experiments were conducted
in water samples collected from Lake Zurich in Switzerland, and from two groundwater treatment facilities in Hanoi, Vietnam, to quantify the effects of matrix composition on As(III) oxidation efficiency and to test the suitability of measured rate constants for modeling As(III) oxidation in real water systems
Materials and Methods
Chemical Reagents As(III) and As(V) stock solutions were
prepared from NaAsO2(purity g99%) and Na2HAsO4‚7H2O (purity g98.5%) obtained from Fluka FAC stock solutions were prepared from NaOCl (∼7% available chlorine) obtained from Riedel-de Hae¨n, and standardized by iodometric
titration (14) Chloramine and O3 stocks were prepared
according to published procedures (15, 16) Additional
reagents were commercially available and of at least reagent grade purity All stock solutions were prepared in deionized water (F g 18.2 MΩ-cm) obtained from a Millipore Milli-Q
or Barnstead NANOpure water purifier Fifty-mM As(III) stocks were prepared approximately bi-monthly, during which time they were stable to within 5% of their initial
* Corresponding author phone: +41 44 823 52 70; fax: +41 44 823
52 10; e-mail: vongunten@eawag.ch
†Swiss Federal Institute of Aquatic Science and Technology
(EAWAG)
‡Hanoi University of Science
§Laboratory of Inorganic Chemistry, ETH Zurich
Environ Sci Technol.2006,40,3285-3292
10.1021/es0524999 CCC: $33.50 2006 American Chemical Society VOL 40, NO 10, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 93285
Trang 2concentration Working As(III) solutions were prepared from
these stocks before each experiment Aluminosilicate
ad-sorbent for the separation of As(III) and As(V) (described in
detail in the Supporting Information, Text S1), was purchased
from Dr Xiaoguang Meng, Stevens Institute of Technology,
Hoboken, NJ
Analytical Methods As(III) and As(V) concentrations were
measured by ion-chromatography/ICP-MS (see Supporting
Information, Text S1 for details) p-Chlorobenzoic acid (pCBA)
and benzaldehyde analyses were performed by HPLC-UV
(Text S1)
Determination of Rate Constants Rate constants for the
reactions of As(III) with FAC, NH2Cl, NHCl2, and O3, were
measured by various techniques selected according to
reactant characteristics and reaction rates In each case,
oxidant consumption was measured in the presence of a
large excess of As(III), to maintain pseudo-first-order
condi-tions with respect to oxidant Individual experimental
procedures are summarized in Table 1 and described in detail
within Text S2
Stoichiometric Measurements Stoichiometries of the
reactions between As(III) and each oxidant (excluding NHCl2)
were measured with either As(III) or the oxidant in excess
In each case, increasing concentrations of the limiting
reactant were dosed under constant, rapid stirring to buffered
solutions of the reactant-in-excess Reactions were allowed
to proceed for time intervals sufficient to ensure complete consumption of the limiting reactant, prior to sampling and analysis for the reactant-in-excess When As(III) was dosed
to solutions containing excess O3, experiments were con-ducted under gastight conditions to minimize evaporative
O3losses (details in Text S3)
Real Water Experiments Water samples used for real
water experiments are listed with corresponding water quality parameters in Table 2 These waters, which contained native
As(III) concentrations <2 µg/L, were spiked with 50 µg/L
As(III) for experiments Separate aliquots of Lake Zurich water were spiked with 1 or 20 mg-N/L (7.1× 10-5or 1.4× 10-3 M) as NH3to simulate waters containing high NH3and low DOC concentrations Sample procurement details and source water descriptions are provided in Text S4 Methods utilized for real water experiments are summarized in Table 1, and described in detail within Text S5
Results and Discussion
As(III) Oxidation Kinetics Pseudo-first-order rate constants,
k′obs, were obtained for reactions of As(III) with each oxidant
by linear regression of plots of ln([Oxidant]) versus time, or
by exponential regression of plots of [Oxidant] versus time, where appropriate (details are provided in the Supporting
TABLE 1 Experimental Approaches Used for Rate Constant Measurements and Real Water Experiments
experiment (method) a T (°C) measurement endpoint e experimental matrix(es) f,g
quenching agent(s) g
NH2Cl kinetics (batchb) 25 ((0.5) NH2Cl loss (measured at λ ) 243 nm) buffered As(III) stock NA
NHCl2kinetics (batchb) 25 ((0.5) NHCl2loss (measured at λ ) 310 nm) buffered As(III) stock NA
FAC kinetics (CFLb) 23 ((2)d FAC loss (measured via DPD method) buffered As(III) stock (C1),
buffered FAC stock (C2)
DPD (C3)
O3kinetics (CFLb) 23 ((2)d O3loss (measured via indigo method) buffered As(III) stock (C1),
pH 4 O3stock (C2)
indigo (C3)
O3kinetics (SFLb) 20 ((0.5) O3loss (measured at λ ) 258 nm) buffered As(III) stock (C1),
buffered O3stock (C2)
NA real water chlorination
(batchc)
25 ((0.5) measurement of As(III) loss for various
FAC doses
As(III)-spiked real water ascorbic acid
or DPDi real water chlorination
(batchc)
25 ((0.5) measurement of As(III) loss in the presence
of various NH2Cl concentrations
As(III)-spiked real water ascorbic acid
or DPDi real water ozonation
(batchc)
20 ((0.5) measurement of As(III) loss for various
O3doses
As(III)-spiked real water none real water chlorination
(CFLc)
23 ((2)d time-resolved monitoring of As(III) loss for
an applied excess of FAC
As(III)-spiked real water (C1), buffered FAC stock (C2)
ascorbic acid
or DPD (C3)i real water ozonation
(CFLc)
23 ((2)d time-resolved monitoring of As(III) loss
for an applied excess of O3
As(III)-spiked real water (C1),
pH 4 O3stock (C2a), buffer (C2b)h
cinnamic acid (C3)j
a CFL-continuous-flow, SFL-stopped-flow b Experimental details included in the Supporting Information, Text S2 c Experimental details included
in the Supporting Information, Text S5 d Room temperature e Reaction kinetics experiments conducted with As(III) in large excess of each oxidant.
Real water experiments conducted at starting concentrations of 50 µg/L As(III) (6.7× 10 -7 M) f Phosphate, acetate, and borate buffers adjusted
to desired pH values in reaction kinetics experiments or, in real water experiments, to the appropriate real water’s native pH O 3 stocks acidified
to pH 4 by dropwise addition of 150 mM sulfuric acid g As(III)-containing solutions introduced on channel 1 (C1) of the CFL system, oxidant solutions on channel 2 (C2), and quenching reagent solutions on channel 3 (C3) h O 3 stock (C2a) and buffer (C2b) premixed to yield a buffered
O 3 stock (C2) at the real water pH immediately prior to further mixing with As(III)-spiked real water (C1) i Ascorbic acid used to quench samples intended for As(III) analyses and DPD used to monitor FAC residuals j The O 3 -cinnamic acid reaction yields benzaldehyde in 1:1 stoichiometry (17) Residual O 3 concentrations were calculated from benzaldehyde concentrations in quenched samples.
TABLE 2 Water Sources and Important Parametric Measurementsa
a Multiple samples of each water, collected on different dates over a four-month time-span, were utilized to conduct the real water experiments described herein; thus, ranges of measurements are provided for each water quality parameter Single values indicate that measurements were the same for each sample b Separate aliquots of native LZ water were amended with NH 4 Cl (LZ1 with 1 mg-N/L (7.1 × 10 -5 M), and LZ20 with
20 mg-N/L (1.4 × 10 -3 M)) for use in batch and time-resolved chlorination experiments.
Trang 3InformationsPart B) As(III) reaction orders were determined
by evaluating the dependence of k′obson [As(III)] for each
reaction Plots of log(k′obs) v log([As(III)]) for FAC, NH2Cl,
and O3reactions all yielded slopes of 1.0 ((0.07) (Figure S1),
indicating that each reaction can be treated as first-order
with respect to As(III) The reactions of As(III) with FAC,
NH2Cl, and O3could thus be described by a second-order
kinetic model (eq 1),
where k′′app(calculated by dividing k′obsby [As(III)])
repre-sents the pH-dependent, apparent second-order rate
con-stant at a particular pH In contrast to its reactions with the
other oxidants, As(III) was found to react with NHCl2with
an order of 0.7 (Figure S1)
Free Available Chlorine Figure 1a, which shows the
magnitude of k′′app,FACat various pH values, illustrates that
As(III) reacts very rapidly with FAC k′′app,FACis 2.6 ((0.1)×
105M-1s-1at pH 7, corresponding to a t1/2,As(III)of 95 ms in
the presence of 2 mg/L Cl2(2.8× 10-5M) as FAC The
pH-dependency of k′′app,FACcan be attributed to varying
contri-butions of each reactant’s acid-base species to apparent
FAC-As(III) reactivity
As(OH)3 dissociates in aqueous solution according to
eqs 2-4 (18).
At circumneutral pH, As(OH)3represents the most abundant
As(III) species A small fraction (up to∼0.1) of As(III) is present
as As(OH)2O-under these conditions, and very small fractions
(<5× 10-6) are present as As(OH)O22-and AsO33-(Figure
1a) FAC speciation can be described by eq 5 (19).
FAC-As(III) reaction kinetics can be characterized according
to eight possible reactions between the four As(III) species
(eqs 2-4) and two FAC species (eq 5), by incorporating species
distribution terms into eq 1 and rearranging to yield eq 6,
where Ri and β jrepresent the respective fractions of oxidant
and substrate present as the species i and j at a given pH
(20), and k′′ij represents the specific second-order rate
constant for each i and j pair.
The increase in magnitude of k′′app,FACup to pH 8.3 can be
attributed primarily to an increase in the fraction of As(III)
present as As(OH)2O-, which is expected to be a stronger
nucleophile than As(OH)3 The decrease in magnitude of
k′′app,FACabove pH 8.3 can be attributed to an accompanying
decrease in proportion of HOCl relative to OCl-, which is a
much weaker oxidant than HOCl (15, 21-23) Consequently,
the magnitude of k′′app,FACis highest near pH 8.3 (the average
of pKa1,As(III)and pKa,HOCl), where the product RHOClβAs(OH) 2 O
-(R1β2) reaches a maximum (Figure 1a) These observations
indicate that OCl-reactions are unimportant relative to HOCl
reactions within the pH range studied Therefore, the
magnitude of k′′app,FACis governed primarily by the reactions
of HOCl with each acid-base species of As(OH)3 k′′app,FAC can thus be modeled by neglecting OCl-reactions
Specific rate constants, k′′1j - calculated by nonlinear
regression of measured k′′app,FACvalues, according to eq 6 (via SigmaPlot 2002, SPSS software), are summarized in Table 3 The model fit shown in Figure 1a, which was obtained by
using these k′′1jvalues, demonstrates the accuracy of eq 6 in
describing measured magnitudes of k′′app,FAC k′′14could not
be accurately determined from available data However, this term is unimportant within the pH range studied, as
HOCl-As(III) reactivity is governed almost exclusively by k′′11, k′′12,
and k′′13under these conditions (Figure 1a)
Chloramines The magnitude of k′′app,NH2Cl is shown at various pH values in Figure 1b These data illustrate that
d([As(III)])
d([Ox])
dt ) - k′′app[As(III)] [Ox] ) -k′obs[Ox] (1)
k′′app,FAC) ∑
i)1,2
j)1,2,3,4
k′′ijR
i [FAC] β j[As(III)]
[FAC] [As(III)]
i)1,2 j)1,2,3,4
k′′ijR
i β j
(6)
FIGURE 1 Apparent second-order rate constants for As(III) oxidation
by (a) FAC, (b) NH 2 Cl, and (c) O 3 FAC experiments conducted at [As(III)] ) 15-50 × 10 -6 M, [FAC] 0 ) 1.5-5 × 10 -6 M, and 23 ((2)
°C, NH 2 Cl experiments at [As(III)] ) 5 × 10 -3 M, [NH 2 Cl] 0 ) 2 ×
10 -4 M, and 25 ((0.5) °C, and O 3 experiments at [As(III)] ) 10-200 × 10 -6 M, [O 3 ] 0 ) 1-10 × 10 -6 M, and 23 ((2)°C (CFL) or
20 ((0.5)°C (SFL).
Trang 4As(III) reacts relatively slowly with NH2Cl k′′app,NH2Cl is 4.3
((1.7)× 10-1M-1s-1at pH 7, corresponding to a t1/2,As(III)of
16 h in the presence of 2 mg/L Cl2(2.8× 10-5M) as NH2Cl
The inverse relationship between pH and magnitude of
k′′app,NH2Clfrom pH 8 to 11 (Figure 1b) is suggestive of
acid-catalysis, in analogy to the reactions of NH2Cl with SO3
2-(24), NO2-(25), and I-(15) This catalysis appears to be H+
-specific, as k′′app,NH2Clexhibited no measurable dependence
on phosphate (50-182 mM) or borate (10-80 mM)
con-centrations
The plateau in magnitude of k′′app,NH2Cl below pH 8
indicates that NH2Cl-As(III) reaction kinetics are not
sig-nificantly influenced by neutral As(OH)3within the pH range
studied, because H+-catalyzed oxidation of As(OH)3would
require that the magnitude of k′′app,NH2Clincrease
continu-ously with increasing acidity The trends in Figure 1b can,
therefore, be attributed to H+-catalyzed reactions of NH2Cl
with one or more anionic As(III) species, according to eq 7,
where k′′′j represents the respective third-order H+
-cata-lysis rate constants for each of the three anionic As(III) species,
j In the context of eq 7, the data in Figure 1b also suggest
that the magnitude of k′′app,NH2Cl is governed primarily by
As(OH)2O- below pH 8 Under these conditions, each
successive unit decrease in pH is offset by an order of
magnitude decrease in the mole fraction of As(OH)2O-,
resulting in a constant value for the product of the [H+] and
β jterms in eq 7 This should, in turn, lead to a constant value
of k′′app,NH2Cl, assuming that As(OH)O22- and AsO33-have
minimal influence on reaction kinetics below pH 8
These inferences were tested by nonlinear regression of
measured k′′app,NH2Clvalues according to eq 7 The resulting
model fit, obtained with the k′′′As(OH)2O -and k′′′As(OH)O22-values
listed in Table 3, is shown in Figure 1b k′′′AsO33-could not be
accurately determined, due to lack of data above pH 11
However, this term is unimportant within the pH range
studied, as the magnitude of k′′app,NH2Clis clearly influenced
primarily by k′′′As(OH)2O -and k′′′As(OH)O22-between pH 6.5 and 11
(Figure 1b)
An Arrhenius plot of k′′app,NH2Clfrom 10 to 30°C showed
that E afor the As(III)-NH2Cl reaction is 27 ((2) kJ/mol
(Supporting Information, Figure S2) A temperature change
of 10°C will, therefore, result in variation of k′′app,NH2Clby a
factor of 1.4-1.5 within temperature ranges relevant to
drinking water treatment
NHCl2-As(III) reaction kinetics were found to be far slower
than NH2Cl-As(III) kinetics k′obs,NHCl2increased from 0.4×
10-5to 2.4× 10-5s-1(i.e., t1/2) 8-48 h) in the presence of
13 mM of As(III), as pH decreased from 4 to 5 (Supporting
Information, Figure S3) These data indicate that the reaction
of As(III) with NHCl2can be neglected under typical drinking
water disinfection conditions
Ozone The magnitude of k′′app,O3, measured by CFL and SFL methods, is shown as a function of pH in Figure 1c As illustrated by these data, As(III) reacts extremely rapidly with
O3 k′′app,O3is 1.5 ((0.1)× 106M-1s-1at pH 7, corresponding
to a t1/2,As(III)of 11 ms in the presence of 2 mg/L O3(4.2× 10-5
M) The magnitude of k′′app,O3is also strongly pH-dependent However, oxidant speciation does not need to be considered for O3reactions, so k′′app,O3can be characterized according to
As(III) speciation alone The constancy of k′′app,O3below pH
6 can be attributed to the O3-As(OH)3reaction, whereas the
increase in k′′app,O3above pH 6 can be attributed primarily to the O3-As(OH)2O-reaction As(OH)O22-and AsO33-exert
negligible influence on the magnitude of k′′app,O3below pH 8.5, since molar fractions of these two species are very small under such conditions (<5× 10-6)
k′′j values were determined for the O3-As(III)
reac-tion by fitting eq 6 to k′′app,O3in the same manner as for the As(III)-FAC reaction (with O3 terms substituted for FAC
terms) The resulting model fit, obtained with the k′′As(OH)3
and k′′As(OH)2O-values listed in Table 3, is shown in Figure 1c
k′′As(OH)O22- and k′′AsO33- could not be accurately determined from available data However, the importance of these terms
is negligible within the pH range studied, as apparent from
the nearly exclusive dependence of k′′app,O3 on k′′As(OH)3 and
k′′As(OH)2O -under these conditions (Figure 1c)
Mechanistic Considerations As mentioned above, the
FAC, NH2Cl, and O3reactions were found to be first-order with respect to As(III) and oxidant (Figure S1) In addition, stoichiometries of As(III) oxidation by FAC, NH2Cl, and O3 were found to be 1:1 for all three reactions; that is, one mole
of As(III) was consumed for each mole of oxidant consumed, whether experiments were conducted with As(III) or oxidant
in excess (Supporting Information, Figure S4) Experiments conducted with As(III) in excess also verified that one mole
of As(V) is produced for every mole of As(III) consumed (Figure S4)
Free Available Chlorine On the basis of reaction order
and stoichiometry, the oxidation of As(OH)3by HOCl, yielding AsO(OH)3, superficially resembles a direct oxygen transfer reaction O-transfer would involve direct nucleophilic sub-stitution by As(III) at the oxygen atom in HOCl, with HCl as
a leaving group However, comparison with FAC reaction systems involving other inorganic nucleophiles (e.g., SO32-,
Br-, I-, CN-(22)) suggests that As(III) oxidation more likely
proceeds via initial Cl+-transfer from HOCl to the As atom, with concomitant loss of OH-(a much more favorable leaving group than HCl), to yield a transient As(III)Cl+intermediate that hydrolyzes to Cl-and As(V) (eqs 8 and 9) This pathway
is expected to apply to HOCl reactions with all four As(III) species
TABLE 3 Specific Rate Constants Determined for Reactions of As(III) with HOCl, NH2Cl, and O3
k′′app(M -1 s -1 ), pH 7 ( 1/2 at 2 mg/L oxidant concentration) b
HOCl As(OH)3 k′′11) 4.3 ((0.8) × 103M-1s-1 2.6 ((0.1)× 105(1/2) 95 ms)
As(OH)2O- k′′12) 5.8 ((0.1) × 107M-1s-1 As(OH)O22- k′′13) 1.4 ((0.1) × 109M-1s-1
NH2Cl As(OH)2O- k2′′′a) 6.9 ((2.7) × 108M-2s-1 4.3 ((1.7)× 10-1(1/2) 16 h)
As(OH)O22- k3′′′a) 8.3 ((7.8) × 1010M-2s-1
O3 As(OH)3 k′′1 ) 5.5 ((0.1) × 105M-1s-1 1.5 ((0.1)× 106(1/2) 11 ms)
As(OH)2O- k′′2 ) 1.5 ((0.1) × 108M-1s-1
a Third-order, H +
-catalysis rate constant b calculated for pseudo-first-order conditions of excess oxidant, assuming 2 mg/L concentrations of
FAC (28 µM), NH2Cl (28 µM), and O3(42 µM).
k′′app,NH2Cl) [H+
j)2,3,4
kHOCl,As(OH)3′′
-(8)
Trang 5Monochloramine NH2Cl is known to react with a number
of inorganic nucleophiles (e.g., SO32-, NO2-, I-) by
acid-catalyzed Cl+-transfer, to yield the same chloro-intermediates
produced in corresponding FAC reactions (15, 24, 25) The
order and 1:1 stoichiometry of the NH2Cl-As(III) reaction,
together with the pH-dependence of k′′app,NH2Cl, are
consis-tent with a similar mechanism (eq 10), which should apply
to oxidation of all three anionic As(III) species The
chloro-intermediate formed in eq 10 would hydrolyze in analogy to
eq 9
Ozone O3generally reacts with inorganic nucleophiles
by two-electron processes involving O-transfer from O3to
the nucleophile via a primary ozonide adduct, which
decomposes to yield the oxidized substrate and O2(26,27).
The order and 1:1 stoichiometry of the As(III)-O3reaction
indicate that As(III) is similarly oxidized to As(V) by O-transfer
from O3to the As atom (eqs 11 and 12) The same pathway
is expected to apply to reactions of O3with all four As(III)
species
Oxidation of As(III) in Real Waters ChlorinationsFree
Available Chlorine Reactions Figure 2a depicts measured
As(III) losses at various FAC doses in each of the real waters
listed in Table 2 Fifty µg/L As(III) (6.7× 10-7M) was
de-pleted to <1 µg/L As(III) by as little as 0.1 mg/L Cl2(1.4×
10-6M) as FAC during batch experiments conducted with LZ
and YP waters (Figure 2a) Such high As(III) oxidation
efficiency is consistent with the low DOC concentrations
and lack of NH3in these two waters (Table 2) In contrast,
As(III) oxidation efficiency was markedly suppressed in
LZ1, LZ20, and PV waters (Figure 2a), due to rapid scavenging
of FAC by the NH3 present in the latter three waters
(k′′app,FAC,NH3> 1 × 104M-1s-1between pH 7 and 8 (23)).
Time-resolved As(III) losses during chlorination of LZ,
LZ1, and YP waters were monitored by the CFL system
mentioned in Table 1 Results obtained from these
experi-ments are summarized in Figure 2b FAC residuals were
present during the monitored reaction periods in all three
waters, ensuring rapid As(III) oxidation in each case The
higher rate of As(III) oxidation in LZ water, compared to YP
water, can be attributed to the difference in pH of the two
waters; with k′′app,FAC,As(III)) 6.9 × 105M-1s-1at pH 8 (LZ
water), and 3.6× 105M-1s-1at pH 7.2 (YP water) Comparison
of the results for LZ and LZ1 waters shows that As(III)
oxidation efficiency was moderately impaired in the latter,
due to rapid consumption of FAC by NH3(Figure 2b) These
findings are consistent with the results obtained for batch
experiments with the same waters (Figure 2a)
The As(III) losses shown in Figure 2b can be modeled
with the rate constants reported in Table 3, by compensating
for contemporaneous FAC loss to side-reactions with matrix
constituents in each water FAC losses were modeled
according to pseudo-first-order rate “constants,” k′FAC,matrix,
obtained from plots of ln([FAC]) vs time in each water
As(III) oxidation was in turn modeled by inserting the
pseudo-first-order expression for FAC decay (eq 13) into a separate
expression for As(III) oxidation (eq 14), and integrating to
yield eq 15
The resulting model fits are shown as dotted lines in Figure 2b The close agreement between model predictions and measured data for each real water demonstrates that one can accurately predict oxidation of As(III) by FAC in various real waters if the rate of FAC loss for a given water is known However, in certain cases, one can make predictions of expected As(III) oxidation efficiencies even without directly measuring FAC loss rates For example, FAC reacts with NH3 far more rapidly than with DOM and most other matrix constituents; thus, in systems containing substantial NH3 concentrations (e.g., >0.5 mg-N/L), FAC loss will likely be dominated by FAC-NH3 reaction kinetics In such cases, FAC loss can be predicted by modeling FAC consumption according to the second-order reaction between NH3and FAC As(III) loss can in turn be modeled by substituting a second-order expression for FAC loss (eq 16) into eq 14 and
integrating with respect to t, as described in the Supporting
Information (Text S6), to yield eq 17
Model predictions obtained by eq 17 are compared with measurements from LZ1 water in Figure 2c The model substantially over-predicted As(III) losses with respect to batch measurements This was presumably a consequence
of suboptimal mixing in the batch systems, which would have resulted in disproportionately large consumption of FAC by NH3during FAC dosage, in turn leading to lower As(III) loss than predicted for an ideally mixed system However, model predictions correlated very well with CFL measurements (Figure 2c), consistent with the superior mixing efficiency achieved by the CFL system (i.e., FAC and real water solutions are mixed through a tee in 1:1 proportion during CFL experiments, as described in the Supporting Information, Text S2)
ChlorinationsNH 2 Cl Reactions As(III) loss is expected to
occur within two phases during chlorination of a water containing significant NH3concentrations: (i) initial, rapid oxidation of As(III) by FAC, and (ii) secondary, slow oxidation
of As(III) in the presence of NH2Cl generated by the
FAC-NH3 reaction, if insufficient FAC is added to completely oxidize As(III) during the first phase As(III) losses measured
kCl′ As(OH)3
-(9)
kH′′′ + ,NH2Cl,As(OH)2O
kOOOAs(OH)3′
[As(III)] ) [As(III)]0e(-kapp,FAC′′ ∫0t[FAC]dt) (14) [As(III)] )
[As(III)]0exp(k′′app,FAC[FAC]0
k′FAC,matrix (e(-kFAC,matrix′ t)-1)) (15)
[FAC] )
[FAC]0(1 -[NH3]0
(1 -[NH3]0 [FAC]0)e(([NH3 ] 0 -[FAC] 0)kapp,FAC,NH3′′ t)
(16)
[As(III)] )
[As(III)]0exp(-k′′app,FAC,As(III)[FAC]0(1 -[NH3]0
([NH3]0-[FAC]0)k′′app,FAC,NH
3
×
(ln( e(([NH3 ] 0 -[FAC] 0)kapp,FAC,NH3 ′′ t)
[NH3]0
(([NH 3 ] 0 -[FAC] 0)kapp,FAC,NH3 ′′ t)-1)- ln( 1
[NH3]0 [FAC]0-1(17)) ))
Trang 6within the latter phase, during chlorination of LZ1, LZ20,
and PV waters, are shown in Figure 2d
With the exception of LZ1 water dosed with 0.1 mg/L Cl2
(1.4× 10-6M) as FAC, NH2Cl concentrations in each water
were in substantial excess of [As(III)], and remained
es-sentially constant during monitored reaction periods in these
waters (i.e., <10% change from [NH2Cl]0, data not shown)
As(III) losses were, therefore, modeled initially by eq 18
However, this model substantially under-predicted the rate
of As(III) loss observed within LZ1, LZ20, and PV waters
(dotted lines in Figure 2d), presumably because it does not
account for effects of the equilibrium between NH2Cl and
HOCl (eq 19) on As(III) oxidation
The importance of eq 19 can be investigated by using the
equilibrium constant, Khyd) 1.5 × 1011M-1(28) to
deter-mine the equilibrium concentration, [HOCl]eq, from known
concentrations of NH3(Table 2) and NH2Cl [FAC]eq(including
both HOCl and OCl-) can then be calculated from [HOCl]eq
and incorporated with k′′app,FACinto eq 18, to yield eq 20, by
which contributions of NH2Cl and HOCl to As(III) loss can
be modeled together
As shown by the solid lines in Figure 2d, eq 20 yielded predictions that are in very good accord with measured As(III) loss in LZ1 and LZ20 waters, illustrating that the
NH2Cl-HOCl equilibrium plays a significant role in governing As(III) loss in the presence of excess NH2Cl However, predictions obtained for PV water by eq 20 still deviated substantially from measured As(III) losses (Figure 2d) The reason for these discrepancies is presently unknown
Ozonation Figure 3a depicts measured As(III) losses for
various O3doses in LZ, YP, and PV waters A dose of only 0.25 mg/L O3(5.2× 10-6M) was sufficient to achieve >99% loss
of 50 µg/L As(III) (6.7 × 10-7 M) in LZ and YP waters
Comparable oxidation of 50 µg/L As(III) was also achieved
in PV water at a relatively low O3 dose (0.8 mg/L O3, or 1.7× 10-5M) (Figure 3a), because O3, in contrast to FAC, reacts very slowly with NH3(k′′app,O3,NH3) 0.2 M-1s-1at pH
7.3 (29)) The observation that more O3than FAC (on a molar basis) is required to achieve comparable oxidation of As(III)
in LZ and YP waters (Figures 2a and 3a) can be attributed
to the higher reactivity of O3toward DOM, which results in comparably more rapid O3loss to side-reactions with water matrix constituents
FIGURE 2 As(III) oxidation during chlorination of real waters spiked with 50 µg/L As(III) (6.7× 10 M) (water quality data in Table 2) (a) As(III) loss 10 s after FAC addition to each real water, in batch at 25 ((0.5)°C, (b) time-resolved As(III) loss within LZ, LZ1, and YP waters for an applied FAC dose of 0.5 mg/L Cl 2 (7.1 × 10 -5 M) at 23 ((2)°C, (c) comparison of As(III) losses measured within LZ1 water,
in batch (25 ((0.5)°C) and by CFL (23 ((2)°C), with As(III) losses predicted for the same water by modeling FAC losses according to the second-order reaction between FAC and NH 3 (d) As(III) losses in LZ1, LZ20, and PV waters, in the presence of NH 2 Cl formed from various FAC doses at 25 ((0.5)°C CC1, CC2, and CC3 represent “combined chlorine” (i.e., NH 2 Cl) concentrations of 0.1, 0.25, and 0.5 mg/L Cl 2 (1.4
× 10 -6 , 3.5 × 10 -6 , and 7.1 × 10 -6 M) for LZ1 water, and 0.5, 1.0, and 1.8 mg/L Cl 2 (7.1 × 10 -6 , 1.4 × 10 -5 , and 2.5 × 10 -5 M) for LZ20 and
PV waters DPD measurements verified that [NH 2 Cl] did not decrease more than 10% during the total reaction times in any of these reaction solutions Symbols in b-d refer to measurements, lines to model predictions.
[As(III)] ) [As(III)]0e(-kapp,NH2Cl′′ [NH2Cl] t) (18)
NH3+ HOCl y\zKhyd
NH2Cl + H2O (19)
[As(III)] ) [As(III)]0e(-(kapp,NH2Cl′′ [NH2Cl]+kapp,FAC ′′ [FAC]eq ) t) (20)
Trang 7Figure 3b, depicting time-resolved measurements of O3
and As(III) losses in LZ, YP, and PV waters, illustrates that
t1/2,O 3is less than 0.33 s in LZ and YP waters, whereas t1/2,FAC
exceeds 1.8 s for the same waters (Figure 2b)
Molar-equivalent doses of FAC and O3, therefore, resulted in similar
rates of As(III) oxidation within these waters, even though
the magnitude of k′′app,O3exceeds that of k′′app,FACby a factor
of 5-15 at circumneutral pH Figure 3b also shows that O3
loss is more rapid in PV water than in LZ or YP waters, due
to the higher DOC concentration in PV water This is
consistent with the comparably lower efficiency of As(III)
oxidation by O3in PV water (Figure 3a) Furthermore, Figure
3b shows that the rate of As(III) loss is significantly faster in
LZ water than in YP water, and approximately equivalent to
the rate of As(III) loss in PV water, even though the latter was
dosed with four times as much O3 This can be attributed in
part to the higher pH of LZ water; that is, k′′app,O3is 9.4× 106
M-1s-1at pH 8 (LZ water), compared to 2.0× 106M-1s-1
at pH 7.2 (YP water) and 2.4× 106 M-1s-1at pH 7.3 (PV
water)
Hydroxyl radicals (‚OH) - generated by autocatalytic
O3decomposition or by direct reactions of O3with water
matrix constituents (30,31) - also react rapidly with As(III)
(k′′‚OH,As(OH)3) 8.5 ((0.9) × 109M-1s-1(32)) p-Chlorobenzoic
acid (pCBA), which reacts rapidly with ‚OH, but is nonreactive
toward O3, was used as an in situ probe (33) to evaluate the
importance of ‚OH-As(III) reactions during ozonation of
each real water The pCBA losses depicted in Figure 3b show
that ‚OH was generated in measurable yield within each
system However, calculated contributions of ‚OH to observed
As(III) losses were very low (i.e., <5% of total observed loss
for LZ and YP waters, and <10% for PV water, see Supporting
Information, Text S7 for a detailed discussion) Time-resolved
measurements of As(III) losses in these waters were, therefore,
modeled by considering only O3-As(III) reaction kinetics
(via eq 15, with O3terms substituted for FAC terms) The
close agreement of model predictions with experimental data
confirms that As(III) loss was dominated by direct reactions
with O3(Figure 3b)
Implications for As(III) Oxidation during Full-Scale
Drinking Water Treatment As demonstrated here and in
prior work (10), oxidant-scavenging matrix constituents such
as NH3and DOM can lower the efficiency of As(III)
preoxi-dation processes Fe(II), which reacts very rapidly with FAC
and O3at pH e 2 (34, 35), may represent another important
oxidant scavenger in such waters, though FAC-Fe(II) and
O3-Fe(II) reaction kinetics must be measured at
circum-neutral pH to permit quantitative evaluation of its potential influence on chlorination or ozonation processes When oxidant-scavenger concentrations are relatively low, their influence on As(III) oxidation efficiency during chlo-rination or ozonation processes will likely be offset by the extremely fast kinetics of FAC-As(III) and O3-As(III) reac-tions However, high scavenger concentrations may sub-stantially impair As(III) oxidation efficiency (Figures 2a and 3a) Proper selection of oxidants can minimize matrix effects
in the latter case For example, ozonation will generally be preferable to chlorination for oxidation of As(III) in waters containing high NH3concentrations (e.g., PV water), because
O3reacts slowly with NH3 In comparison, chlorination is likely to prove more efficient than ozonation for As(III) oxidation in waters lacking NH3, because FAC typically reacts more slowly than O3with DOM over time-scales relevant to FAC-As(III) reactions (Figures 2b and 3b)
In waters with high oxidant scavenging rates, As(III) oxidation efficiencies will also be highly sensitive to mixing efficiency during oxidant application (Figure 2c) The high sensitivity of FAC-As(III) and O3-As(III) reaction kinetics to
pH (Figure 1) indicates that pH control may also play an important role in As(III) oxidation efficiency Careful attention
to these considerations will facilitate optimization of oxidant dose when As(III) oxidation must be balanced with con-straints such as disinfection byproduct formation
In an optimized chlorination or ozonation process, complete preoxidation of As(III) should generally be achiev-able at oxidant doses for which disinfection byproduct formation will be minimal For example, THM and NDMA formation potentials in YP and PV waters are known to be far below WHO, EU, and USEPA limits at the FAC doses required to achieve full As(III) oxidation within these waters
during the present investigation (13) Bromate formation
during ozonation of these waters is also expected to be low, because YP water contains low Br-concentrations (i.e., <30
µg/L), and PV water contains high NH3concentrations, which will substantially suppress bromate formation by scavenging HOBr generated by reaction of O3with Br-(36).
NH2Cl formed during chlorination of ammoniacal waters will likely only have appreciable effect on As(III) fate in special cases; for example, if source waters undergo limited or no treatment prior to chlorination, and insufficient FAC is added
to directly oxidize As(III) during chlorination Although the direct NH2Cl-As(III) reaction may result in minimal As(III) oxidation after chlorination, indirect NH2Cl-mediated oxida-tion reacoxida-tions can yield substantial As(III) oxidaoxida-tion within
FIGURE 3 As(III) oxidation during ozonation of real waters spiked with 50 µg/L As(III) (6.7× 10 M) (water quality data in Table 2) (a) As(III) loss for various O 3 doses within real waters included in this study, in batch at 25 ((0.5)°C, (b) time-resolved As(III) loss within
LZ water ([O 3 ] 0 ) 0.25 mg/L (5.2 × 10 -6 M), 23 ((2)°C), YP water ([O 3 ] 0 ) 0.25 mg/L (5.2 × 10 -6 M), 23 ((2)°C), and PV water ([O 3 ] 0)
1 mg/L (2.1 × 10 -5 M), 23 ((2)°C) Symbols refer to measurements, lines to model predictions.
Trang 8such systems over reaction times of several hours (e.g., within
disinfection contact chambers or distribution networks), as
illustrated in Figure 2d
Acknowledgments
M.C.D and N.D.V contributed equally to this work Travel
scholarships and financial support for N.D.V and V.C.L were
obtained from the Swiss Agency for Development and
Cooperation (SDC), in the framework of the Swiss-Vietnamese
project ESTNV (Environmental Science and Technology in
Northern Vietnam) M.C.D gratefully acknowledges financial
support from a U.S National Science Foundation Graduate
Research Fellowship The authors thank Elisabeth Salhi,
Caroline Stengel, and Sebastien Meylan for their technical
assistance Willem Koppenol is acknowledged for support in
obtaining stopped-flow measurements of As(III)-O3reaction
kinetics The authors also thank Stephan Hug, Linda Roberts,
Olivier Leupin, Marc-Olivier Buffle, and Gretchen Onstad
for many helpful discussions The Hanoi Water Works
Company is acknowledged for assistance in obtaining water
samples from Hanoi
Supporting Information Available
Tables and figures addressing experimental methods and
modeling approaches, water sample sources and
procure-ment, reaction orders and stoichiometries, As(III)-NHCl2
reaction kinetics, and temperature-dependence of
As(III)-NH2Cl reaction kinetics, in addition to reaction kinetics data
from which rate constants were determined This material
is available free of charge via the Internet at http://
pubs.acs.org
Literature Cited
(1) Cullen, W R.; Reimer, K J Arsenic Speciation in the
Environ-ment Chem Rev 1989, 89, 713-764.
(2) Frey, M M.; Edwards, M A Surveying arsenic occurrence J.
Am Water Works Ass 1997, 89, 105-117.
(3) Berg, M.; Tran, H C.; Nguyen, T C.; Pham, H V.; Schertenleib,
R.; Giger, W Arsenic contamination of groundwater and drinking
water in Vietnam: A human health threat Environ Sci Technol.
2001, 35, 2621-2626.
(4) Smith, A H.; Lingas, E O.; Rahman, M Contamination of
drinking-water by arsenic in Bangladesh: A public health
emergency Bull W H O 2000, 78, 1093-1103.
(5) Guidelines for Drinking-water Quality; 3rd ed.; World Health
Organization: Geneva, Switzerland, 2004; Vol 1
(6) Hering, J G.; Chiu, V Q Arsenic occurrence and speciation in
municipal groundwater-based supply system J Environ Eng.,
ASCE 2000, 126, 471-474.
(7) Council Directive 98/83/EC of 3 November 1998 on the quality
of water intended for human consumption Off J Eur
Com-munities 1998, 41, 32-54.
(8) National Primary Drinking Water Regulations Code of Federal
Regulations, Part 141, Title 40, 2001, http://www.epa.gov/
epahome/cfr40.htm
(9) Technologies and Costs for Removal of Arsenic from Drinking
Water, United States Environmental Protection AgencysOffice
of Water: Washington, DC, 2000 (http://www.epa.gov/
safewater/ars/treatments_and_costs.pdf)
(10) Ghurye, G.; Clifford, D As(III) oxidation using chemical and
solid-phase oxidants J Am Water Works Assoc 2004, 96,
84-96
(11) Leupin, O X.; Hug, S J.; Badruzzaman, A B M Arsenic removal
from Bangladesh tube well water with filter columns containing
zerovalent iron filings and sand Environ Sci Technol 2005,
39, 8032-8037.
(12) McArthur, J M.; Ravenscroft, P.; Safiulla, S.; Thirlwall, M F.,
Arsenic in groundwater: Testing pollution mechanisms for
sedimentary aquifers in Bangladesh Water Resour Res 2001,
37, 109-117.
(13) Duong, H A.; Berg, M.; Hoang, M H.; Pham, H V.; Gallard, H.;
Giger, W.; von Gunten, U Trihalomethane formation by
chlorination of ammonium- and bromide-containing
ground-water in ground-water supplies of Hanoi, Vietnam Water Res 2003, 37,
3242-3252
(14) Standard Methods for the Examination of Water and Wastewater,
20th ed.; APHA, AWWA, WPCF: Washington DC, 1998 (15) Kumar, K.; Day, R A.; Margerum, D W Atom-transfer redox kinetics: General-acid-assisted oxidation of iodide by
chlor-amines and hypochlorite Inorg Chem 1986, 25, 4344-4350.
(16) Bader, H.; Hoigne´, J Determination of ozone in water by the
indigo method Water Res 1981, 15, 449-456.
(17) Leitzke, A.; Reisz, E.; Flyunt, R.; von Sonntag, C The reactions
of ozone with cinnamic acids: formation and decay of
2-hy-droperoxy-2-hydroxyacetic acid J Chem Soc., Perkin Trans 2
2001, 793-797.
(18) Sadiq, M.; Zaidi, T H.; Mian, A A Environmental behavior of
arsenic in soilsstheoretical Water Air Soil Pollut 1983, 20,
369-377
(19) Lide, D R., Ed CRC Handbook of Chemistry and Physics, 82nd
ed.; CRC Press: Boca Raton, FL, 2001
(20) Stumm, W.; Morgan, J J Aquatic Chemistry, 3rd ed.; John Wiley
and Sons: New York, 1996
(21) Fogelman, K D.; Walker, D M.; Margerum, D W Non-metal redox kinetics: hypochlorite and hypochlorous acid reactions
with sulfite Inorg Chem 1989, 28, 986-993.
(22) Gerritsen, C M.; Margerum, D W Non-metal redox kinetics: hypochlorite and hypochlorous acid reactions with cyanide
Inorg Chem 1990, 29, 2757-2762.
(23) Qiang, Z.; Adams, C D Determination of monochloramine formation rate constants with stopped-flow spectrophotometry
Environ Sci Technol 2004, 38, 1435-1444.
(24) Yiin, B S.; Walker, D M.; Margerum, D W Nonmetal redox kinetics: general-acid-assisted reactions of chloramine with
sulfite and hydrogen sulfite Inorg Chem 1987, 26, 3435-3441.
(25) Margerum, D W.; Schurter, L M.; Hobson, J.; Moore, E E Water Chlorination Chemistry: Nonmetal Redox Kinetics of
Chlor-amine and Nitrite Ion Environ Sci Technol 1994, 28,
331-337
(26) Hoigne´, J In The Handbook of Environmental Chemistry; Hrubec,
J., Ed.; Springer-Verlag: Berlin, Germany, 1998; pp 83-141 (27) Liu, Q.; Schurter, L M.; Muller, C E.; Aloisio, S.; Francisco, J S.; Margerum, D W Kinetics and mechanisms of aqueous ozone reactions with bromide, sulfite, hydrogen sulfite, iodide, and
nitrite ions Inorg Chem 2001, 40, 4436-4442.
(28) Gray, E T., Jr.; Margerum, D W.; Huffman, R P In Organometals and Organometalloids, Occurrence and Fate in the Environment;
Brinkman, F E., Bellama, J M., Eds.; American Chemical Society: Washington, D C., 1978; pp 264-277
(29) Hoigne´, J.; Bader, H Ozonation of water: Kinetics of oxidation
of ammonia by ozone and hydroxyl radicals Environ Sci.
Technol 1978, 12, 79-84.
(30) Buffle, M.-O.; Schumacher, J.; Salhi, E.; von Gunten, U Measurement of the initial phase of ozone decomposition in water and wastewater by means of a continuous quench flow system: Application to disinfection and pharmaceutical
oxida-tion Water Res 2006, in press.
(31) von Gunten, U Ozonation of drinking water: Part I Oxidation
kinetics and product formation Water Res 2003, 37,
1443-1467
(32) Klaening, U K.; Bielski, B H J.; Sehested, K Arsenic(IV)sa
pulse-radiolysis study Inorg Chem 1989, 28, 2717-2724.
(33) Elovitz, M S.; von Gunten, U Hydroxyl radical/ozone ratios during ozonation processes I The Rctconcept Ozone: Sci Eng.
1999, 21, 239-260.
(34) Conocchioli, T J.; Hamilton, E J.; Sutin, N Formation of
Iron-(IV) in Oxidation of Iron(II) J Am Chem Soc 1965, 87,
926-927
(35) Løgager, T.; Holcman, J.; Sehested, K.; Pedersen, T Oxidation
of Ferrous-Ions by Ozone in Acidic Solutions Inorg Chem 1992,
31, 3523-3529.
(36) Pinkernell, U.; von Gunten, U Bromate minimization during
ozonation: Mechanistic considerations Environ Sci Technol.
2001, 35, 2525-2531.
Received for review December 13, 2005 Revised manuscript received March 7, 2006 Accepted March 10, 2006.
ES0524999